The Day the CPK Report Was Released, Silence Fell for Three Seconds, and the Boss's Face Went Pale
Do you remember a few years ago? We had an old machine then. Even though the process parameters were identical, the yield rate fluctuated significantly, experiencing unusually low performance on several days each month. Once, the CPK report came out, dropping directly to 1.08, and DPMO soared to 6210. The entire production line was stunned. The boss's face turned green during the report meeting and he immediately ordered us to find the problem. And what do you guess? After a full day of investigation, it turned out not to be a process issue at all, but rather repetitive errors caused by aging equipment and worn-out parts. Only then did everyone realize that we were so focused on pushing production capacity that we completely neglected these old partners!
Where Was the Problem? Simply Put, It Was "Treating the Symptom, Not the Root Cause"
Many times, we only fix machines when the yield drops or the machine breaks down. This is a typical case of "reactive maintenance." You might think you're being efficient, but in reality, every unexpected incident disrupts your production schedule, leading to delayed deliveries and even requiring more time and resources to resolve. This is like your car: waiting until it won't start to fix it—isn't that more expensive and troublesome? In a semiconductor factory, every time equipment goes down, it's like burning piles of cash. Therefore, what we need to do is to transform passive "repair" into proactive "maintenance." This is the core spirit of TPM (Total Productive Maintenance).
How Is It Actually Done? Let Operators Become "Machine Doctors"
Simply put, TPM is "Total Productive Maintenance." It's not just the responsibility of equipment engineers; even operators must get involved. You might think, what do operators know about these things? But that's precisely the point! They spend the most time with the machines daily and best understand the machines' "temperament."
- Autonomous Maintenance: At that time, we started teaching operators to perform basic inspections, cleaning, lubrication, and tightening screws. For example, checking vacuum levels before starting up each day and cleaning residues in the chamber. Don't underestimate these small actions; when a machine makes an unusual sound, they can immediately detect it, instead of waiting until the machine completely breaks down. Later, we found that simply by operators spending 10 minutes on daily checks, equipment abnormal events decreased by 15%!
- Planned Maintenance: This is the responsibility of us equipment engineers. Based on the machine's historical data, we develop preventive maintenance plans. For example, if a certain part has an average lifespan of 5000 hours, we schedule its replacement at 4500 hours. This is much simpler than waiting for the part to fail before replacing it.
- Quality Maintenance: This part ensures stable quality after equipment maintenance. After each part replacement, we perform a mini-run to ensure process stability; for instance, the Cpk value must return to above 1.33 to be considered acceptable.
The Most Common Pitfall: Thinking That Implementing TPM Means "Increasing Workload"
The most common resistance encountered when promoting TPM is the perception that "there's yet another new task." Initially, operators found machine cleaning troublesome, and equipment engineers felt that training others was a waste of time. Another time, we asked operators to record some parameters, but they carelessly scribbled them down, making the data unusable. Frankly, this was all because people hadn't seen the "value of doing these things." We later changed our strategy, converting the abnormal events reduced by TPM into "saved overtime costs" and "increased bonuses," and even posted charts directly on the production line showing "how much TPM improved yield." When everyone saw these tangible benefits, motivation naturally followed.
One Thing You Can Do Today
Observe the machine that most frequently experiences issues and write down three of its most common abnormal phenomena.